Model LineUpper: Supporting Interactive Model Comparison at Multiple Levels for AutoML

TL;DR

Model LineUpper is a tool that enhances AutoML by enabling interactive, multi-level model comparisons using Explainable AI techniques, addressing the gap in current systems that rely solely on performance metrics.

Contribution

It introduces a novel interactive system that integrates multiple XAI and visualization methods for comprehensive model comparison in AutoML.

Findings

Users found the tool improved understanding of models.

The system facilitated more informed model selection.

Design implications for XAI in AutoML were identified.

Abstract

Automated Machine Learning (AutoML) is a rapidly growing set of technologies that automate the model development pipeline by searching model space and generating candidate models. A critical, final step of AutoML is human selection of a final model from dozens of candidates. In current AutoML systems, selection is supported only by performance metrics. Prior work has shown that in practice, people evaluate ML models based on additional criteria, such as the way a model makes predictions. Comparison may happen at multiple levels, from types of errors, to feature importance, to how the model makes predictions of specific instances. We developed \tool{} to support interactive model comparison for AutoML by integrating multiple Explainable AI (XAI) and visualization techniques. We conducted a user study in which we both evaluated the system and used it as a technology probe to understand how users perform model comparison in an AutoML system. We discuss design implications for utilizing XAI techniques for model comparison and supporting the unique needs of data scientists in comparing AutoML models.